Vol. 23, No. 2INFECTION AND IMMUNITY, Feb. 1979, p. 522-5310019-9567/79/02-0522/10$02.00/0

Myeloperoxidase, Hydrogen Peroxide, Chloride AntimicrobialSystem: Nitrogen-Chlorine Derivatives of Bacterial

Components in Bactericidal Action Against Escherichia coliEDWIN L. THOMAS

Department of Biochemistry, St. Jude Children's Research Hospital and The University of TennesseeCenter for the Health Sciences, Memphis, Tennessee 38101

Received for publication 6 November 1978

In the presence of Escherichia coli, myeloperoxidase-catalyzed oxidation ofchloride ion resulted in formation of long-lived chloramine and/or chloramidederivatives of bacterial components. The same amount of these nitrogen-chlorine(N-Cl) derivatives was obtained with either hypochlorous acid (HOCl) or themyeloperoxidase system, indicating that myeloperoxidase catalyzed the oxidationof chloride to HOCl. Identical killing was obtained with HOCI or the myelope-roxidase system. About 30 to 50% of the oxidizing equivalents of HOCl weredetected as N-Cl derivatives of peptides or peptide fragments that were releasedfrom the bacteria. The apparent molecular weight distribution of the peptidesdecreased with increasing amounts of HOCl, suggesting that peptides werefragmented by oxidative cleavage of chloramide derivatives of peptide bonds. Theremaining 50 to 70% of the oxidizing equivalents of HOCI were rapidly consumedin peptide bond cleavage or the oxidation of other bacterial components. Therewas a close correspondence between the oxidation of bacterial sulfhydryls andbactericidal action. The N-Cl derivatives were lost and the oxidation of bacterialsulfhydryls increased over a period of several h at 37°C. These changes were

accompanied by increased killing. The increase in sulfhydryl oxidation and killingcould be prevented by washing the bacteria to remove the N-Cl derivatives.Therefore, the N-Cl derivatives could oxidize bacterial components long after themyeloperoxidase-catalyzed oxidation of chloride was complete.

The combination of myeloperoxidase, hydro-gen peroxide (H202), and chloride ion (Cl-)forms an antimicrobial system within phagocyticvesicles of monocytes and polymorphonuclearleukocytes (4, 5, 20, 21, 30, 33, 38, 44). Undersome conditions, myeloperoxidase and H202 canbe released from leukocytes (6, 7). This releasemay permit antimicrobial activity outside theleukocytes, but may also result in damage tohost tissues. The myeloperoxidase system hascytotoxic activity against both normal and tu-mor cells, and may contribute to inflammatorytissue destruction (11, 22).Agner (1) first proposed that myeloperoxidase

catalyzes the oxidation of Cl- by H202 to yieldhypochlorous acid, HOCl.

H202 + Cl- + H+ myeloperoxidase HOCl + H20

Recently, the oxidation of Cl- to HOCl wasdemonstrated by using a flow system that con-tinuously removed HOCl, so as to prevent inac-tivation of myeloperoxidase by HOCI (14). An-other problem in detecting HOCI formation is

the reaction of HOCI with H202. Although H202acts as an oxidizing agent in the generation ofHOCI, it acts as a reducing agent in the reactionwith HOCl. The net result of Cl- oxidation andHOCI reduction is the same as catalyzed by theenzyme catalase, resulting in the loss of theoxidizing equivalents of H202 and the formationof molecular oxygen, 02.

2 H202 myeloperoxidase, C1- 2 H20 + 0

Although the reduction of HOCI by H202would appear to eliminate any antimicrobial ac-tivity, a reactive species may be formed as anintermediate in this reaction. Under alkalineconditions, chlorine (Cl2) reacts with a deproton-ated form of H202 to yield singlet oxygen (02'),which is an activated form of 02 (9)-

Cl2 + H02- + OH- - 02' + 2 Cl + H20

The O2' is very short-lived and rearranges withthe emission of light to yield 02.

021 -- 02+ hv

522

N-Cl DERIVATIVES OF BACTERIAL COMPONENTS 523

Incubation of myeloperoxidase with H202 andCl- at neutral or acid pH may yield O21 (29, 30),although the evidence for detection of 02' hasbeen questioned (15, 16). Either HOCI (1, 20) or02' (20, 29, 31) has been proposed to be theantimicrobial agent produced by the myelope-roxidase system.The two proposed antimicrobial mechanisms

differ in a way that might permit experimentaldiscrimination. If all the HOCI is reduced to 021,then no long-lived oxidizing agents would bedetected upon incubating the myeloperoxidasesystem with bacteria. If HOCI is the antimicro-bial agent, it might be possible to detect theaccumulation of oxidizing equivalents, assumingthat HOCl is not instantly consumed in theoxidation of bacterial components.Accumulation of oxidizing equivalents has

been observed in the myeloperoxidase-catalyzedformation of chloramines (37,47). The deproton-ated form of amines reacts with HOCI to yieldderivatives containing the N-Cl bond.

R-NH2 + HOCl -- R-NH-CI + H20In the presence of amines, myeloperoxidase-cat-alyzed oxidation of Cl- yielded about 1 mol ofchloramine derivatives per mol of H202, indicat-ing that the reaction of HOCl with amines pro-tected myeloperoxidase against inactivation (37,47). This result also suggests that the chlora-mines were not reduced by H202.Other nitrogenous compounds can also react

with HOCl, such as in the conversion of amidesto chloramides (17).

R-CO-NH-R' + HOCl-- R-CO-NHCl-R' + H20

Chloramines and chloramides retain the oxidiz-ing equivalents of HOCl and are powerful oxi-dizing agents. On the other hand, some of thesederivatives undergo internal oxidation, whichconsumes the oxidizing equivalents ofHOCl. Forexample, chloramine derivatives of a-aminoacids undergo oxidative decarboxylation to yieldCl-, carbon dioxide, ammonia, and an aldehyde(39, 49, 50). Similarly, incubation of proteinswith sodium hypochlorite (NaOCl) results inpeptide bond cleavage (8), which suggests inter-nal oxidation of chloramide derivatives of theamide nitrogens of peptide bonds.

Incubation with the myeloperoxidase systemresulted in release of radioactivity from bacteriagrown in the presence of [14C]diaminopimelicacid (34). This observation may indicate decar-boxylation of diaminopimelic acid residues withfree carboxyl groups, or peptide bond cleavage.Another study showed that myeloperoxidase

catalyzed the incorporation of Cl- into bacterialcomponents (48). However, that study did notdistinguish between formation of N-Cl deriva-tives (chloramines and chloramides) and C-Clderivatives (chlorinated aromatic or aliphaticcompounds). Unlike the N-Cl derivatives, the C-Cl derivatives do not act as oxidizing agents.The aims of the present study were to deter-

mine whether incubation of bacteria with themyeloperoxidase system resulted in accumula-tion of oxidizing equivalents as N-Cl derivativesof bacterial components and to determinewhether such derivatives have any role in bac-tericidal action. In addition, the reaction ofHOCl with amines or amides may prevent thereduction of HOCl by H202 so that detection ofN-Cl derivatives could rule out the participationof 02' in myeloperoxidase-catalyzed bactericidalaction.

MATERIALS AND METHODS

Chemicals. All chemicals were of reagent grade.Myeloperoxidase purified from human leukemic gran-ulocytes was provided by M. Morrison and J. Naskal-ski. Dithiothreitol and 5,5'-dithiobis(2-nitrobenzoicacid) were obtained from Sigma Chemical Co., St.Louis, Mo., lysozyme was from Worthington Biochem-icals Corp., Freehold, N.J., poly-L-lysine hydrochloride(average molecular weight, 375,000) was fromSchwarz/Mann, Orangeburg, N.Y., and NaOCl wasfrom Fisher Scientific Co., Fair Lawn, N.J. Sodiumborohydride was used to reduce 5,5'-dithiobis(2-nitro-benzoic acid) to 5-thio-2-nitrobenzoic acid (3).

Bacteria and isolated bacterial components.Escherichia coli ML 308-225 was grown aerobically at37°C in medium A (10), modified by addition of 5 mMtris(hydroxymethyl)aminomethane (Tris)-Cl (pH 6.6)with 1% (wt/vol) disodium succinate as the carbonsource. Bacteria were harvested in the late logarithmicphase of growth by centrifugation at 5,300 x g for 10min.

Bacteria were extracted by the method of Leive (25)to obtain components released from the outer mem-brane by extraction with ethylenediaminetetraaceticacid (EDTA). Bacteria were subjected to the osmoticshock procedure of Neu and Heppel (27) to obtain thesoluble "periplasmic" proteins. To obtain outer andinner membrane vesicles, bacteria were treated withlysozyme and EDTA by the method of Osborn et al.(28) and subjected to osmotic lysis and low-speedcentrifugation by the method of Kaback (18), and thenthe supernatant was fractionated by sucrose density-gradient centrifugation to yield low-density (innermembrane) and high-density (outer membrane) frac-tions by a modification of the method of Osborn et al.(28), in which the gradient solutions were bufferedwith 0.1 M phosphate buffer, pH 6.6. Soluble cyto-plasmic proteins were obtained by the sonic oscillationmethod of Ames (2), by using osmotically shockedbacteria. The soluble fractions were dialyzed againstwater, clarified by centrifugation, and lyophilized.

VOL. 23, 1979

524 THOMAS

Membrane fractions were collected by centrifugationand suspended in water. Protein concentration wasdetermined by the method of Lowry et al. (26), withbovine serum albumin as the protein standard.Exposure of bacteria to the myeloperoxidase

system or HOCI. Bacteria were suspended to 6 x 108cells per ml in 0.2 M NaCl, with 0.1 M potassiumphosphate buffer (pH 6.6) and 1 mM MgSO4 at 0-50C.The suspensions were made to 0.1 [LM in myeloperox-idase. At 1-min intervals, H202 was added, with eachaddition sufficient to give a concentration of 30 /LM.Incubations were continued after the last addition ofH202 for a total of 30 min of incubation. Alternatively,HOCl was generated by adding portions of 20 mMNaOCl in 0.1 M sodium hydroxide to the bufferedbacterial suspension. The NaOCl was added as a singleaddition, and the incubation was continued for 30 minat 0 to 50C. All reactions that were carried out withoutbacteria were performed in the same 0.2 M NaCl-0.1M phosphate buffer, but without MgSO4.

Quantitation of HOCI and N-Cl derivatives.The concentration of HOCl or N-Cl derivatives wasdetermined by the oxidation of 2 mol of the sulfhydrylcompound, 5-thio-2-nitrobenzoic acid, to the disulfidecompound, 5,5'-dithiobis(2-nitrobenzoic acid), by 1mol of the oxidizing agent. An excess of the sulfhydrylcompound was added to the reaction mixtures, andabsorbance at 412 nm was measured. Centrifugationat 5,300 x g for 5 min at 0 to 5VC was used to removebacteria. A molar extinction coefficient (E) of 13,600was assumed for 5-thio-2-nitrobenzoic acid (12).

Sulflydryl determinations. Two-milliliter por-tions of reaction mixtures were diluted with 6 ml ofthe buffer with MgSO4, then centrifuged at 5,300 x gfor 10 min at 0 to 5°C. The bacteria were resuspendedin 2 ml of the same buffer.

Sulfhydryls were measured by the reaction of 1 molof sulfhydryl with 5,5'-dithiobis(2-nitrobenzoic acid) toyield 1 mol of 5-thio-2-nitrobenzoic acid (12). Theresuspended bacteria were diluted with 4 ml of 0.1 MTris-Cl buffer (pH 8) with 10 mM EDTA. Portions of0.1 ml of 10 mM 5,5'-dithiobis(2-nitrobenzoic acid) in0.1 M phosphate buffer (pH 7), and 0.5 ml of 10% (wt/vol) sodium dodecyl sulfate were added. The mixtureswere incubated for 1 h at 37°C, cooled to 0 to 5°C, andthen centrifuged at 18,000 x g for 10 min. Absorbanceof the supernatants was measured at 412 nm.

Electrophoresis. Samples were solubilized in 40mM Tris-Cl buffer (pH 6.8), containing 2.1 M urea,2.5% (wt/vol) sodium dodecyl sulfate, 3.5% (vol/vol)2-mercaptoethanol, and 0.0025% (wt/vol) bromo-phenol blue. Peptides were separated according tomolecular weight by the method of Laemmli (24), byusing a 7.5 to 15% (wt/vol) gradient of acrylamide.Peptides were visualized by staining with Coomassiebrilliant blue (13).Assay of viability. Serial 1:10 dilutions were pre-

pared from 1-ml portions of reaction mixtures in sterile0.2 M NaCl with 0.1 M phosphate buffer (pH 6.6) and1 mM MgSO4. One-milliliter portions of appropriatedilutions were spread on a solid medium containingmodified medium A, 2% (wt/vol) agar (Difco Labora-tories, Detroit, Mich.) and 1% glucose. Plates wereincubated at 25°C, and the number of colonies wererecorded after 2 to 3 days.

RESULTSFormation of N-Cl derivatives. Figure 1

shows the yield of oxidizing equivalents obtainedeither as N-Cl derivatives or as HOCl after in-cubation of the myeloperoxidase system withammonium ion (NH4+), or without other addi-tions. In the presence of NH4', about 1 mol of aN-Cl derivative was obtained per mol of H202,consistent with myeloperoxidase-catalyzed oxi-dation of Cl- to yield 1 mol of HOCl per mol ofH202, followed by the reaction of HOCl withNH4'. The product was assumed to be mono-chloramine (NH2CI), rather than NHC12 or NC13,based on the high ratio of NH4' to HOCl.About 1 mol of N-Cl derivatives was obtained

per mol of H202 with all the primary and sec-ondary amines tested, at pH 5, 6.6, or 7.4. Theyields were lower in the presence of the tertiaryamine trimethylamine, or the amide compoundsacetamide and N-acetylglycine, or at pH 8.

In the absence of amines or amides, there wasa very low yield of HOCl. After this incubation,additions of NH4', myeloperoxidase, and H202were required to obtain NH2Cl. Therefore, allthe myeloperoxidase had been inactivated andall the H202 had been consumed in the oxidationof Cl- and the reduction of HOCl. When theoxidation of Cl- was carried out with a 10-fold

300

E

200

-1000I

0 00 200 300H202 (nmol)

FIG. 1. Myeloperoxidase-catalyzed formation ofHOCI or N-Cl derivatives. Reaction mixtures con-tained 0.1 pM myeloperoxidase and the indicatedamounts of H202 per ml in the 0.2 M NaCl-0.1 Mphosphate buffer with no other additions (V), with 5mM (NH4)2SO4 (O), with bacteria and 1 mM MgSO4(0), or with the supernatant obtained by suspendingbacteria in the buffer with MgSO4 and then centrifug-ing at 5,300 X g for 10 min at 0 to 5°C (A). Incubationswere continued for a total of 30 min at 0 to 5°C, andthen the amounts of HOCI or N-Cl derivatives weremeasured.

INFECT. IMMUN.

N-Cl DERIVATIVES OF BACTERIAL COMPONENTS 525

lower myeloperoxidase concentration, no HOCIwas obtained and excess unreacted H202 re-mained. Therefore, the myeloperoxidase hadbeen inactivated before an amount of HOCI wasgenerated equal to half the amount of H202added.

Figure 1 also shows that in the presence of E.coli accumulation of oxidizing equivalents wasobserved. Therefore, the E. coli suspension con-tained a sufficient amount of amines and/oramides to convert most or all of the HOCIformed to N-Cl derivatives. Lowering the celldensity by a factor of 10 had little effect on theyield of N-Cl derivatives, indicating that theyield of these derivatives was not limited by theamount of amines and/or amides. On the otherhand, the yield of N-Cl derivatives was less than1 mol per mol of H202. Therefore, part of theoxidizing equivalents of HOCl had been con-sumed in the oxidation of bacterial components.

Figure 1 also shows that the supernatant ob-tained by centrifuging the bacterial suspensiondid not support the accumulation of oxidizingequivalents. Therefore, the amines and/oramides involved in formation of N-Cl derivativeswere in or on the bacteria and not in the sus-pending fluid.

Figure 2 shows that unreacted oxidizing equiv-alents were also detected when HOC1 was gen-erated by adding a solution of NaOCl to thebuffered E. coli suspension. The yield of N-Clderivatives was nearly identical to that obtainedwith the myeloperoxidase system. These resultsconfirm that the low yield of N-Cl derivativesobtained with the myeloperoxidase system wasnot due to reduction of HOCI by H202. Instead,part of the HOC1 was consumed in the oxidationof bacterial components. About 50 to 70% of theHOC1 added was consumed.

Figure 2 also shows the amount of N-Cl deriv-atives detected either in the supernatant or as-sociated with the bacteria after centrifugation.Although results presented above indicated thatformation of N-Cl derivatives was due to reac-tion of HOCl with bacterial components, mostof the N-Cl derivatives were found in the super-natant after centrifugation. As described below,amine and/or amide components were releasedfrom the bacteria as N-Cl derivatives.The yield of N-Cl derivatives obtained with

HOCl or the myeloperoxidase system was aboutthe same at 37°C or at 0 to 50C. On the otherhand, the N-Cl derivatives were stable at 0 to5°C, but were slowly lost during prolonged in-cubation at 370C. About 40% of the oxidizingequivalents remained after 2 h at 37°C. In theexperiments reported here, bacteria were ex-posed to the myeloperoxidase system at 0 to 5°Cto improve the discrimination between fast re-

200E

100

&

0 200 400 600HOCI (nmol)

FIG. 2. Yield ofN-Cl derivatives from HOCI. Bac-teria were incubated 30 min at 0 to 5°C with theindicated amounts of HOCI per ml. The amount ofN-Cl derivatives was measured in the total reactionmixtures (0). Also, reaction mixtures were centri-fuged at 5,300 x g for 10 min at 0 to 5°C, and then theamounts of N-Cl derivatives were measured in thesupernatant (A) and in the suspensions obtained byresuspending the bacteria to the original volume(V).

actions (such as C1- oxidation and formation ofN-Cl derivatives) and slow reactions (such asloss of N-Cl derivatives).The rate of loss of the N-Cl derivatives at

37°C was about the same either in the completereaction mixture or in the supernatant aftercentrifugation to remove the bacteria. There-fore, the loss of N-Cl derivatives was not pri-marily due to oxidation of other bacterial com-ponents by the N-Cl derivatives. Internal oxi-dation of the N-Cl derivatives may have contrib-uted to the loss.Differing reactivity of HOCI and N-Cl de-

rivatives. Figure 3 shows that the addition ofH202 to HOC1 resulted in reduction of 1 mol ofHOC1 per mol of H202. On the other hand, theoxidizing equivalents were not lost when NH4'(or an amine or amide) was present, indicatingthat the N-Cl derivatives were not reduced byH202. The same results were obtained whenHOC1 was added to solutions containing H202and NH4'. Therefore, with 10 mM NH4' and 0.1to 0.3 mM H202, the rate of formation of N-Clderivatives was much faster than the rate ofreduction of HOC1.

Oxidation of 5-thio-2-nitrobenzoic acid byH202 was negligible during the time required forthe measurements shown in Fig. 3. Also, un-reacted H202 was detected after the incubationof N-Cl derivatives with H202.The difference in reactivity of HOCl and N-Cl

derivatives could be used to demonstrate for-mation of N-Cl derivatives of bacterial compo-nents. Figure 3 shows that in the presence of E.coli, H202 did not reduce the unreacted oxidizing

VOL. 23, 1979

526 THOMAS

E_S ---- 'O --

150

03

Z100

I\

50

0 50 100 150 200H202 (nmol)

FIG. 3. Differing reactivity of HOCI and N-Cl de-rivatives with H202. The indicated amounts of H202per ml were added to reaction mixtures containing0.25 mM HOCI (0), 0.25mM HOCI and 10 mM NH4'(a), bacteria that had been incubated with 0.6 mMHOCI for 30 min at 0 to 50C (A), or the supernatantobtained by incubating bacteria with 0.6 mM HOCIfor 30 min at 0 to 5VC and then centrifuging at 5,300x g for 10 min at 0 to 5VC (V). After the addition ofH202, the incubations were continued for 15 min at 0to 5VC, and then the amounts ofHOCI or N-Cl deriv-atives were measured.

equivalents. The same results were obtainedwhen H202 was added to the supernatant frombacteria that had been incubated with HOC.Therefore, the oxidizing equivalents were in theform of N-Cl derivatives, rather than HOCl.Released N-Cl derivatives. Incubation of

E. coli with HOCl resulted in release of proteinin a soluble form. Upon addition of 0.1, 0.4, and1.2 mM HOCl, the amounts of protein obtainedin the supernatant were 50, 65, and 90 ,ig/ml,respectively, after centrifugation to remove bac-teria. These amounts of protein were less than6% of the dry weight of the bacteria. The amountof released material increased up to twofoldupon prolonging the incubation for 2 h at 370C.The supernatant fractions were treated with

dithiothreitol to eliminate oxidizing agents, di-alyzed against water to remove salts, and thenconcentrated by lyophilization. Figure 4Athrough H shows that the protein fractions pre-pared in this way could be resolved by electro-phoresis into a number of peptides with appar-ent molecular weights of about 100,000 to 10,000.In other experiments, a similar distribution ofpeptides was obtained by incubating the bacteriawith the myeloperoxidase system. Consistent

with the results of others (19), myeloperoxidasewas bound to the bacteria. Myeloperoxidase didnot appear in the supernatant either when addedalone or with H202.The amounts of peptides observed on gels

increased upon prolonged exposure of the bac-teria to HOCL. On the other hand, the amountsappeared to decrease with increasing amounts ofHOCl. As observed on gels, the largest amountof peptides was obtained by treating E. coli withabout 0.1 mM HOCI, whereas direct proteindeterminations on the supernatants indicated

rn

A B C D E F G H

-~~tm-

K ~~~IV N

FIG. 4. Separation of released peptides by electro-phoresis. The samples contained: supernatant frac-tions from bacteria incubated with 0, 0.1, 0.4, and 1.2mM HOCI for 30 min at 0 to 50C (A through D) or foran additional incubation period of 2 h at 37C (Ethrough H); the supernatant fraction from bacteriasubjected to osmotic shock (I and L); supernatantfractions from bacteria incubated with 0.1 m.MHOCIfor 30 min at 0 to 50C (J) or for an additionalincubation period of2 h at 370C (K); the supernatantfraction from bacteria extracted with EDTA (M);outer membrane (N); inner membrane (0); cytoplas-mic proteins (P). Bacteria were removed from reac-tion mixtures treated with HOCI by centrifugation at18,000 x g for 20 min at 0 to 50C, and the supernatantfractions were reduced with 1 mM dithiothreitol, di-alyzed against water, and lyophilized.

INFECT. IMMUN.

Allen

N-Cl DERIVATIVES OF BACTERIAL COMPONENTS 527

increasing protein concentrations with increas-ing amounts of HOCL. These results suggestedthat the released peptides were fragmented byHOCi. Peptides of molecular weight less than8,000 would have been lost during the dialysisstep before electrophoresis.

Figure 4I through P shows that the apparentmolecular weight distribution of the HOCl-re-leased peptides was similar to that of the solubleperiplasmic proteins. The permeability of theouter membrane may have been increased byexposure to HOCl, resulting in release of solubleproteins from the space between the outer andinner membranes. Also, some of the releasedmaterial may have been derived from outer-membrane components in that componentscharacteristic of the outer membrane were de-tected by staining the gel for carbohydrate. Fig-ure 4I through P also shows that the distributionof the released peptides differed from that ofother isolated cell fractions, including the outer-membrane components that are released byEDTA extraction, the outer membrane, the in-ner membrane, and the cytoplasmic proteins.This observation suggested that release of pep-tides was not due to lysis of the cells, whichwould release the soluble proteins, or to solubil-ization of the cell membranes, which would re-lease all the proteins.To demonstrate that the N-Cl derivatives ob-

tained in the supernatant were derivatives of thereleased peptides or peptide fragments, a pro-cedure was developed for the chromatographicseparation of the nonreduced N-Cl derivativeson Sephadex G-50, followed by titration of theseparated derivatives with 5-thio-2-nitrobenzoicacid. Figure 5 shows that high- and low-molec-ular-weight N-Cl derivatives were sufficientlystable to be separated and quantitated in thisway. When polylysine was mixed with HOCl andthen chromatographed, a high-molecular-weight(excluded) fraction was obtained, correspondingto derivatives of polylysine containing multipleN-Cl bonds.

Dissociation of the poly(N-Cl)-polylysine toyield HOCl did not appear significant within thetime required for chromatography (ca. 15 to 30min). Although a low-molecular-weight fractionwith oxidizing activity was obtained, this activitywas not due to HOCl. Other experiments indi-cated that HOCl did not emerge from the col-umn; no oxidizing agent was detected in thecolumn effluent when the sample consisted ofHOCI added to buffer without polylysine orother amines. A low-molecular-weight fractioncorresponding to HOCl could be obtained onlyif the column material was pretreated with ex-cess NaOCl. Apparently, HOCI was consumedin oxidation of the column material, whereas the

81

CS 41

>

2(

o< ~~~~~B.0~~ ~

- C. .yk~~CDo D _

~o4(

160

120

80

40

0

80

40

0

40

0

40 60 80 100 120 140EFFWENT (mL)

FIG. 5. Chromatographic fractionation ofN-Cl de-rivatives. Reaction mixtures containing 0.6mMHOCIand various amounts ofpolylysine were incubated for15 min at 0 to 5°C. The reaction mixtures were 10mM(A), 5 mM (B), 1 mM (C), and 0.2 mM (D) in lysineresidues. Alternatively, the reaction mixture con-sisted of bacteria incubated with 0.3mMHOCI for 30min at 0 to 5°C. The mixture was centrifuged at 5,300x g for 10 min at 0 to 5°C to yield a supernatantfraction (E). Also, a reaction mixture (10 ml) consist-ing of bacteria incubated with 0.3 mM HOCI for 30min at 0 to 5°C was centrifuged, and the supernatantwas reduced with 1 mM dithiothreitol, dialyzedagainst water, lyophilized, suspended in 2 ml of the0.2 M NaCl-0.1 M phosphate buffer, and then incu-bated with 0.3 mM HOCI for 15 min at 0 to 5°C (F).Fractions A through F were chromatographed on acolumn (60 by 1.6 cm) ofSephadex G-50 (PharmaciaFine Chemicals) and equilibrated and eluted with the0.2 M NaCl-0.1 M phosphate buffer. Fractions ofabout 2 ml were collected, the volumes were deter-mined by weighing, and then the amounts of N-Clderivatives were measured.

N-Cl derivatives did not react with the columnmaterial within the time required for chroma-tography. With the reaction mixture 10 mM inlysine residues, the recovery of oxidizing equiv-alents in the effluent was 93% of that applied tothe column.

Also, the low-molecular-weight fraction ob-tained in the presence of polylysine was not dueto formation of N-Cl derivatives of low-molecu-lar-weight contaminants. The same results wereobtained with polylysine that had been dialyzedor pre-chromatographed on Sephadex G-50

VOL. 23, 1979

528 THOMAS

to remove low-molecular-weight substances.Therefore, the low-molecular-weight fractionappeared to be due to fragmentation of thepolylysine and formation of N-Cl derivatives ofthe fragments.

Figure 5 shows that when the concentrationof lysine residues relative to HOCI was lowered,the amount of high-molecular-weight N-Cl de-rivatives decreased, whereas the low-molecular-weight derivatives increased in amount and de-creased in average size. With molar ratios oflysine residues to HOCl of 8.3 and 1.7, the recov-ery of oxidizing equivalents was 93 and 83%,respectively. With a ratio of 0.3, no high-molec-ular-weight derivatives were detected, and therecovery fell to 11%. Therefore, low ratios ofamine groups to HOCl favored fragmentation,possibly by oxidative cleavage of amide bonds.In other experiments, similar results were ob-tained with protein substrates in place of poly-lysine.

Figure 5 also shows that only a low-molecular-weight oxidizing fraction was obtained when thesupernatant from HOCl-treated E. coli waschromatographed. Therefore, the oxidizingequivalents were associated with peptide frag-ments or other low-molecular-weight sub-stances. The recovery of oxidizing equivalentswas 36% of that added to the bacteria and 91%of that applied to the column.

In other experiments, the supernatant wasreduced with dithiothreitol, dialyzed to removelow-molecular-weight substances, incubatedagain with HOCl, and then chromatographed.Only low-molecular-weight N-Cl derivativeswere obtained, indicating that fragmentation ofpeptides was the probable source of low-molec-ular-weight N-Cl derivatives. Figure 5 showsthat when a larger volume of the supernatantwas reduced, dialyzed, concentrated by lyophi-lization, incubated with HOCl, and then chro-matographed, both high- and low-molecular-weight derivatives were obtained. Therefore, theinability to demonstrate high-molecular-weightderivatives in the original supernatant was dueto the low protein concentration, which favoredfragmentation.Oxidation of sulfhydryls. Figure 6 shows

the oxidation of E. coli sulfhydryls by HOCl.Similar results were obtained with the myelo-peroxidase system, consistent with myeloperox-idase-catalyzed oxidation of Cl- to HOCl. About1 mol of sulfhydryls was oxidized per 3.7 mol ofHOCl consumed, assuming that 60% of the HOClwas consumed (as in Fig. 2). This result mayindicate that most of the HOCl was consumedin the oxidation of components other thansulfhydryls. Alternatively, the sulfhydryls may

80 1

~700-60

E

IX Xcr) 30-

20

10 ~ A

I00 _ l__l_0 200 400 600

HOCI (nmol)

FIG. 6. Oxidation of sulfhydryls. Bacteria were in-cubated for 30 min at 0 to 50C with the indicatedamounts of HOCi per ml. Portions were taken fordetermination of the amount of sulfhydryls per ml ofreaction mixtures (0). The remainders were incu-bated 2 h at 370C, and then the sulfhydryls weredetermined (A).

have been oxidized to sulfonic acids, whichwould require 3 mol of HOCl per mol of sulfhy-dryls.

Figure 6 also shows that the amount of oxi-dized sulfhydryls increased when the bacteriawere incubated in the presence of the N-Cl de-rivatives for 2 h at 370C. This increased sulfhy-dryl oxidation could be prevented by washingthe bacteria before the incubation at 370C or byholding the bacteria at 0 to 5VC. Therefore, theN-Cl derivatives were responsible for the in-creased oxidation.

Bactericidal action. Figure 7 shows the kill-ing of E. coli by either HOCl or the myeloperox-idase system. Killing is expressed as the loga-rithm of the ratio of viable cells in an untreatedsuspension to that in a suspension exposed toHOCl or the myeloperoxidase system. A valueof zero indicates a ratio of 1 or no loss of viability,whereas a value of 4 indicates that only 1 cell in10,000 remained viable.

Identical results were obtained when killingper mol of HOCl was compared to killing permol of the H202 added to the myeloperoxidasesystem. These results were consistent with my-eloperoxidase-catalyzed oxidation of Cl- to yield1 mol of HOCl per mol of H202. The amount ofHOCl required for complete killing was aboutthe same as required for complete oxidation ofsulfhydryls (cf. Fig. 6).

INFECT. IMMUN.

N-Cl DERIVATIVES OF BACTERIAL COMPONENTS 529

LUJ

Ld

-

C)LUJ

z

0

HOCI OR H202 (nmol)FIG. 7. Bactericidal action. Bacteria were incu-

bated with the indicated amounts ofHOCI (, A) perml of reaction mixture or with 0.1 ,uM myeloperoxi-dase and the indicated amounts of H202 (0, A) perml. After a 30-min incubation at 0 to 5°C, portionswere taken for determination of viability (0, 0). Theremainders were incubated for 2 h at 37°C, and thenviability was measured (A, A).

Removing the N-Cl derivatives by washingthe bacteria or by reduction with dithiothreitolhad no effect on the loss of viability as measuredimmediately after the 30-min incubation period.On the other hand, Fig. 7 shows that continuedincubation of the bacteria in the presence of theN-Cl derivatives for 2 h at 37°C resulted inincreased killing. This increase could be pre-vented by washing the bacteria, by adding di-thiothreitol, or by holding the bacteria at 0 to5°C for 2 h. Therefore, the N-Cl derivativesappeared responsible for the .increased killingobtained during prolonged incubations.

DISCUSSIONRole of HOCI and 02'. Killing of E. coli by

the myeloperoxidase system was due to the ox-idation of C1- to HOCl, followed by the reactionof HOCl with bacterial components. The amineand/or amide components of the bacteria were

present in sufficient amounts to convert all theHOCI to N-Cl derivatives. The reaction of HOClwith amines or amides competes with and thusprevents the reaction of HOCl with H202, whichmay yield 02'. Therefore, it would appear thatno 021 was produced in the presence of thebacteria and that killing was due only to HOCL.Consistent with this interpretation, identicalkilling was obtained either with the myeloper-

oxidase system or with non-enzymatically gen-erated HOCI.The presence of bacteria alters the chemistry

of the myeloperoxidase system. The chemicalconstituents of the bacteria can suppress O21formation, and also probably protect myeloper-oxidase against inactivation. Amine and/oramide components of the bacteria can be consid-ered as a fourth component of the myeloperoxi-dase system. Exogenous amines can also partic-ipate in killing (E. L. Thomas, unpublisheddata).The involvement of bacterial components pro-

vides an explanation for results of experimentsin which the myeloperoxidase system is physi-cally separated from the bacteria by a dialysismembrane. No killing is obtained under theseconditions (32). The probable explanation is thatthe HOC1 produced in the myeloperoxidase com-partment is reduced by H202 before it can diffuseinto the bacterial compartment. Although 021may be produced in the myeloperoxidase com-partment, 02' would be too short-lived to diffuseinto the bacterial compartment. Consistent withthis interpretation, killing can be obtained whenNH4' or amines are added to the myeloperoxi-dase compartment (33).Because the conditions within phagocytic ves-

icles of leukocytes would be different from thoseused in this study, it is not possible to concludethat reduction of HOCl by H202 does not occurwithin leukocytes. However, amines and amidesare present in significant amounts within phag-ocytic vesicles. The phagocytized bacteriummakes up a large part of the volume of thephagocytic vesicle so that the bacterial compo-nents would be present in high concentrations.Also, leukocytes contain high concentrations ofthe amine compound taurine (36) and arginine-rich cationic proteins that are found within thephagocytic vesicle (45, 46). Bacterial and leuko-cyte components could react with HOCI andprevent O21 formation.Sulfhydryl oxidation. A close correspond-

ence was observed between the oxidation ofbacterial sulfhydryls and bactericidal action. Ox-idation of sulfhydryls has previously been pro-posed as the mechanism of antibacterial actionof C12 (23). Nevertheless, results presented heredo not indicate that sulfhydryl oxidation is theonly reaction that contributes to killing. Theamount of oxidizing equivalents that was con-sumed was much greater than the amount ofoxidized sulfhydryls. Other reactions that maycontribute to the consumption of oxidizingequivalents include oxidation of other easily ox-idized bacterial components, chlorination of bac-terial components to yield C-Cl derivatives, and

VOL. 23, 1979

530 THOMAS

internal oxidation of N-Cl derivatives.In part, the consumption of oxidizing equiva-

lents may have been due to the oxidation ofsulfhydryls to sulfonic acids, which would con-sume 3 mol of HOCl per mol of sulfhydryls. Theability of HOCl and N-Cl derivatives to oxidizesulfhydryls (R-SH) was the basis for their quan-titation.

R-NH-Cl + 2 R-SH-* R-NH3+ + R-S-S-R + Cl-

Although 5-thio-2-nitrobenzoic acid was oxi-dized to the disulfide compound when presentin excess, this and other sulfhydryl compoundscan be oxidized to sulfonic acids by excess HOCl(35).Only about 2% of the oxidized bacterial sulfhy-

dryls were in the form of sulfenyl derivatives(Thomas, unpublished data), indicating thatsulfhydryls were oxidized to other forms such assulfonic acids. These results differ from thoseobtained with the peroxidase, H202, I- or thio-cyanate (SCN-) systems (41, 42). With I- orSCN- as the cofactor, a major portion of thebacterial sulfhydryls were oxidized to sulfenylderivatives. Exogenous reducing agents couldpartly reverse the antibacterial action of thesesystems (40-43), apparently by reducing the sul-fenyl derivatives back to sulfhydryls. In contrast,reducing agents did not reverse the antimicro-bial action of HOCl.Role ofN-Cl derivatives. Internal oxidation

of N-Cl derivatives appeared responsible for thefragmentation of peptides to yield low-molecu-lar-weight N-Cl derivatives. This reaction mayalso have contributed to the release of the pep-tides from the bacterial cell envelope. Thebreaking of peptide or other amide bonds couldcontribute to killing by disrupting bacterialstructure and inactivating essential enzymes. Onthe other hand, killing can be obtained in thepresence of NH4+ or exogenous amines, eventhough they compete with bacterial componentsfor formation of N-Cl derivatives and diminishthe release of peptides (Thomas, unpublisheddata). Also, disruption of the outer membrane orthe loss of the periplasmic proteins does notresult in loss of viability of E. coli (25, 27).The stable N-Cl derivatives appeared to have

a role in killing during incubations of severalhours at 370C. Oxidation of bacterial sulfhydrylsand killing continued to increase, although allthe oxidizing equivalents were in the form of N-Cl derivatives. Removal or reduction of the N-Cl derivatives prevented the increase in oxida-tion and killing. The bacterial sulfhydryls mayhave been oxidized by HOCl obtained from thehydrolysis of N-Cl derivatives.

R-NH-Cl + H20 R-NH2 + HOCl

However, the extent of formation of HOCl fromN-Cl derivatives appeared insignificant even un-der conditions that should shift this equilibrium,such as the separation of the derivatives fromHOCl by chromatography, or the addition ofH202. It is more likely that the slow killing wasdue to direct oxidation of bacterial componentsby the N-Cl derivatives. The relative abilities ofN-Cl derivatives to oxidize bacterial componentswill be considered in another report.

ACKNOWLEDGMENTS

We thank M. Morrison and G. Schonbaum for helpfuldiscussions and Kate Pera Bates and M. Margaret Jeffersonfor technical assistance.

This work was supported by Public Health Service researchgrant DE 04235 from the National Institute of Dental Re-search and Cancer Center Support (CORE) grants CA 08480and CA 21765 from the National Cancer Institute, and byALSAC.

LITERATURE CITED1. Agner, K. 1972. Biological effects of hypochlorous acid

formed by "MPO"-peroxidation in the presence of chlo-ride ions, p. 329-335. In A. Akeson and A. Ehrenberg(ed.), Structure and function of oxidation-reduction en-zymes. Pergamon Press, Inc., Oxford.

2. Ames, G. F.-L. 1974. Resolution of bacterial proteins bypolyacrylamide gel electrophoresis on slabs. J. Biol.Chem. 249:634-644.

3. Aune, T. M., and E. L. Thomas. 1977. Accumulation ofhypothiocyanite ion during peroxidase-catalyzed oxi-dation of thiocyanate ion. Eur. J. Biochem. 80:209-214.

4. Baboir, B. M. 1978. Oxygen-dependent microbial killingby phagocytes. N. Engl. J. Med. 298:659-668, 721-725.

5. Baehner, R. L. 1974. Molecular basis for functional dis-orders of phagocytes. J. Pediatr. 84:317-327.

6. Baehner, R. L., M. J. Karnovsky, and M. L. Karnov-sky. 1969. Degranulation of leukocytes in chronic gran-ulomatous disease. J. Clin. Invest. 48:187-192.

7. Baehner, R. L., D. G. Nathan, and W. B. Castle. 1971.Oxidant injury of caucasian glucose-6-phosphate dehy-drogenase-deficient red blood cells by phagocytizingleukocytes during infection. J. Clin. Invest. 50:2466-2473.

8. Baker, R. W. R. 1947. Studies on the reaction betweensodium hypochlorite and proteins. Biochem. J. 41:337-342.

9. Browne, R. J., and E. A. Ogryzlo. 1964. Chemilumines-cence from the reaction of chlorine with aqueous hy-drogen peroxide. Proc. Chem. Soc. 1964:117.

10. Davis, B. D., and E. S. Mingioli. 1950. Mutants ofEscherichia coli requiring methionine or vitamin B12. J.Bacteriol. 60:17-28.

11. Edelson, P. J., and Z. A. Cohn. 1973. Peroxidase-me-diated mammalian cell cytotoxicity. J. Exp. Med. 138:318-323.

12. Ellman, G. L. 1959. Tissue sulfhydryl groups. Arch. Bio-chem. Biophys. 82:70-77.

13. Fairbanks, G., T. L. Steck, and D. F. H. Wallach. 1971.Electrophoretic analysis of the major polypeptides ofthe human erythrocyte membrane. Biochemistry 10:2606-2617.

14. Harrison, J. E., and J. Schultz. 1976. Studies on thechlorinating activity of myeloperoxidase. J. Biol. Chem.251:1371-1374.

15. Harrison, J. E., B. D. Watson, and J. Schultz. 1978.

INFECT. IMMUN.

N-Cl DERIVATIVES OF BACTERIAL COMPONENTS

Myeloperoxidase and singlet oxygen: a reappraisal.FEBS Lett. 92:327-332.

16. Held, A. M., and J. K. Hurst. 1978. Ambiguity associatedwith use of singlet oxygen trapping agents in myelope-roxidase-catalyzed oxidations. Biochem. Biophys. Res.Commun. 81:878-885.

17. Johnson, R. A., and F. D. Green. 1975. Chlorinationwith N-chloro amides. I. Inter- and intramolecular chlo-rination. J. Org. Chem. 40:2186-2192.

18. Kaback, H. R. 1971. Bacterial membranes. Methods En-zymol. 22:99-120.

19. Klebanoff, S. J. 1970. Myeloperoxidase-mediated anti-microbial systems and their role in leukocyte function,p. 89-110. In J. Schultz (ed.), Biochemistry of the phag-ocytic process. North-Holland, Amsterdam.

20. Klebanoff, S. J. 1975. Antimicrobial mechanisms in neu-trophilic polymorphonuclear leukocytes. Semin. He-matol. 12:117-142.

21. Klebanoff, S. J. 1975. Antimicrobial systems of the pol-ymorphonuclear leukocyte, p. 45-59. In J. A. Bellantiand D. H. Dayton (ed.), The phagocytic cell in hostresistance. Raven Press, New York.

22. Klebanoff, S. J., R. A. Clark, and H. Rosen. 1976.Myeloperoxidase-mediated cytotoxicity. Cancer En-zymol. 12:267-285.

23. Knox, W. E., P. K. Stumpf, D. E. Green, and V. H.Auerbach. 1948. The inhibition of sulfhydryl enzymesas the basis of the bactericidal action of chlorine. J.Bacteriol. 55:451-458.

24. Laemmli, U. K. 1970. Cleavage of structural proteinsduring the assembly of the head of bacteriophage T4.Nature (London) 227:680-685.

25. Leive, L. 1965. Release of lipopolysaccharide by EDTAtreatment of E. coli. Biochem. Biophys. Res. Commun.21:290-296.

26. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1951. Protein measurement with the Folinphenol reagent. J. Biol. Chem. 193:265-275.

27. Neu, H. C., and L. A. Heppel. 1965. The release ofenzymes from Escherichia coli by osmotic shock andduring the formation of spheroplasts. J. Biol. Chem.240:3685-3692.

28. Osborn, M. J., J. E. Gander, E. Parisi, and J. Carson.1972. Mechanism of assembly of the outer membraneof Salmonella typhimurium. Isolation and characteri-zation of cytoplasmic and outer membrane. J. Biol.Chem. 247:3962-3972.

29. Piatt, J. F., A. S. Cheema, and P. J. O'Brien. 1977.Peroxidase-catalyzed singlet oxygen formation from hy-drogen peroxide. FEBS Lett. 74:251-254.

30. Quie, P. G., E. L. Mills, and B. Holmes. 1977. Molecularevents during phagocytosis by human neutrophils. Prog.Hematol. 10:193-210.

31. Rosen, H., and S. J. Klebanoff. 1977. Formation ofsinglet oxygen by the myeloperoxidase-mediated anti-microbial system. J. Biol. Chem. 252:4803-4810.

32. Sbarra, A. J., B. B. Paul, A. A. Jacobs, R. R. Strauss,and G. W. Mitchell, Jr. 1972. Role of the phagocyte inhost-parasite interactions. XXXVIII. Metabolic activi-ties of the phagocyte as related to antimicrobial action.RES J. Reticuloendothel. Soc. 12:109-126.

33. Sbarra, A. J., R. J. Selvaraj, B. B. Paul, P. K. F.Poskitt, G. W. Mitchell, Jr., F. Louis, and M. A.Asbell. 1977. Granulocyte biochemistry and a hydrogenperoxide-dependent antimicrobial system, p. 29-48. InT. J. Greenwalt and G. A. Jamieson (ed.), The granu-

locyte: function and clinical utilization. Alan R. Liss,Inc., New York.

34. Selvaraj, R. J., B. B. Paul, R. R. Strauss, A. A. Jacobs,and A. J. Sbarra. 1974. Oxidative peptide cleavageand decarboxylation by the MPO-H202-C- antimicro-bial system. Infect. Immun. 9:255-260.

35. Silverstein, R. M., and L. P. Hager. 1974. The chloro-peroxidase-catalyzed oxidation of thiols and disulfidesto sulfenyl chlorides. Biochemistry 13:5069-5073.

36. Soupart, P. 1962. Free amino acids in blood and urine inthe human, p. 220-262. In J. T. Holden (ed.), Aminoacid pools. Elsevier, Amsterdam.

37. Stelmaszynska, T., and J. M. Zgliczynski. 1974. Mye-loperoxidase of human neutrophilic granulocytes aschlorinating enzyme. Eur. J. Biochem. 45:305-312.

38. Stossal, T. P. 1974. Phagocytosis. New Engl. J. Med. 290:717-723, 774-780, 833-839.

39. Strauss, R. R., B. B. Paul, A. A. Jacobs, and A. J.Sbarra. 1971. Role of the phagocyte in host-parasiteinteractions. XXVII. Myeloperoxidase-H202-Cl -me-diated aldehyde formation and its relationship to anti-microbial activity. Infect. Immun. 3:595-602.

40. Thomas, E. L., and T. M. Aune. 1978. Cofactor role ofiodide in peroxidase antimicrobial action against Esch-erichia coli. Antimicrob. Agents Chemother. 13:1000-1005.

41. Thomas, E. L., and T. M. Aune. 1978. Lactoperoxidase,peroxide, thiocyanate antimicrobial system: correlationof sulfhydryl oxidation with antimicrobial action. Infect.Immun. 20:456-463.

42. Thomas, E. L, and T. M. Aune. 1978. Oxidation ofEscherichia coli sulfhydryl components by the peroxi-dase-hydrogen peroxide-iodide antimicrobial system.Antimicrob. Agents Chemother. 13:1006-1010.

43. Thomas, E. L., and T. M. Aune. 1978. Susceptibility ofEscherichia coli to bactericidal action of lactoperoxi-dase, peroxide, and iodide or thiocyanate. Antimicrob.Agents Chemother. 13:261-265.

44. Winkelstein, J. A., and R. H. Drachman. 1974. Phag-ocytosis. The normal process and its clinically signifi-cant abnormalities. Pediatr. Clin. N. Am. 21:551-569.

45. Zeya, H. I., and J. K. Spitznagel. 1966. Cationic proteinsof polymorphonuclear leukocyte lysozomes. I. Resolu-tion of antibacterial and enzymatic activities. J. Bacte-riol. 91:750-754.

46. Zeya, H. I., and J. K. Spitznagel. 1966. Cationic proteinsof polymorphonuclear leukocyte lysozomes. II. Com-position, properties, and mechanisms of antibacterialaction. J. Bacteriol. 91:755-762.

47. Zgliczynski, J. M., R. J. Selvaraj, B. B. Paul, T.Stelmaszynska, P. K. F. Poskitt, and A. J. Sbarra.1977. Chlorination by the myeloperoxidase-H202-Cl-antimicrobial system at acid and neutral pH. Proc. Soc.Exp. Biol. Med. 154:418-422.

48. Zgliczynski, J. M., and T. Stelmaszynska. 1975. Chlo-rinating ability ofhuman phagocytizing leukocytes. Eur.J. Biochem. 56:157-162.

49. Zgliczynski, J. M., T. Stelmaszynska, J. Domanski,and W. Ostrowski. 1971. Chloramines as intermedi-ates of oxidation reaction of amino acids by myelope-roxidase. Biochim. Biophys. Acta. 235:419-424.

50. Zgliczynski, J. M., T. Stelmaszynska, W. Ostrowski,J. Naskalski, and J. Sznajd. 1968. Myeloperoxidaseof human leukaemic leucocytes. Oxidation of aminoacids in the presence of hydrogen peroxide. Eur. J.Biochem. 4:540-547.

531VOL. 23, 1979